----------------------------- **While I generate an answer, here are some sources that might be helpful**: * [Update Assistant]() * [Get Assistant]() * [List Assistants]() * [Web Calling]() -----------------------------

### Answer The behavior you're describing may stem from a misconfiguration in how the assistant's models or messages are managed during the conversation. Ideally, the assistant should not vocalize placeholder or internal notes. If this behavior is occurring, you might consider reviewing the configurations related to the assistant's speech generation settings. Specifically, you should check the settings for the assistant's model and messages to ensure they are correctly structured to prevent unintended vocalizations. This could involve adjusting how the assistant handles silence or its response triggers. Unfortunately, specific details on adjusting these settings are not available in the documents I can access, so I recommend reviewing the assistant's configuration in your Vapi dashboard or consulting the API documentation related to the assistant's model settings [(1)]() . If the issue persists and is not addressed by configuration adjustments, consider reaching out to Vapi support for a more tailored solution.

Got it. Tagging @User for additional help.

or fine-tune your groq llm model.